Darwinism, Design, and Complex Systems Dynamics

P1: KAF/KAA P2: KaF
0521829496c09.xml CY335B/Dembski 0 521 82949 6 March 10, 2004 1:5
9
Darwinism, Design, and Complex Systems Dynamics
Bruce H. Weber and David J. Depew
1.
the argument from incredulity
When Saint Augustine was a young man he could not imagine how evil
could possibly have come into the world. Later, he came to realize that this
line of questioning was dangerous and could bring religion into disrepute
(Confessions VI, Chapter 5; VII, Chapters 3–5). His was a wise realization.
‘How possibly’ types of questions may do effective rhetorical work in the early
stages of a line of inquiry or argumentation, but the types of answers they en-
gender are open to refutation by the answers to other types of questions, such
as ‘why actually’ or ‘why necessarily’ questions. ‘Why actually’ questions –
or at least an important class of them that are asked by scientists – have a
remarkable way of shutting down ‘how possibly’ questions.
When looking at the functional complexity of the living world, ‘how pos-
sibly’ questions emerge rather intuitively. As Michael Ruse has pointed out,
the traditional argument from design actually has two steps (Ruse 2002).
The first move is from observed functional complexity to the notion that
such phenomena appear to have been designed, the argument to design.
During this process, alternative, natural explanations are rejected using the
criterion of incredulity: how possibly could the vertebrate eye appear as a
result of random events, even under natural processes and laws? This was
the line of argument taken by William Paley some two centuries ago (Paley
1802). In effect, he dared anyone to come up with a fully natural explana-
tion of biological functional complexity. The remainder of the project of
natural theology is to argue from design to the existence of a designer, and
even to deduce attributes of the designer God from the properties of His
designed Creation.

Darwin, who as an undergraduate had read and admired Paley, took
Paley’s dare and made his research program that of finding a natural ex-
planation of apparent design by way of biological adaptation. Darwin’s ex-
planatory mechanism was natural selection acting upon the random but
173
P1: KAF/KAA P2: KaF
0521829496c09.xml CY335B/Dembski 0 521 82949 6 March 10, 2004 1:5
174
Bruce H. Weber and David J. Depew
heritable variations of reproducing organisms in particular environments.
This selective process produced, over generational time, adaptive traits in
lineages, as well as diversification – descent with modification. Much of
Darwin’s “long argument” was to show that natural selection could account
for the empirical claim of a common descent for all living beings.
1
Many
of Darwin’s sympathizers, John Stuart Mill among them, thought the ar-
gument no stronger than that (Mill 1874, 328). But Darwin was not just
responding to the ‘how possibly’ question posed by Paley; he was also pro-
viding a conceptual framework within which ‘why actually’ questions could
be coherently formulated and answered. The Darwinian research tradition
developed through a succession of research programs over the course of the
twentieth century. Under the aegis of rubrics such as genetical Darwinism or
the Modern Evolutionary Synthesis, questions about adaptation, geograph-
ical distribution, speciation, and related matters were asked and answered
with ever greater empirical and theoretical richness. The Synthesis may not
have been a complete theory of evolution, but it has resolved enough ques-
tions to create a presumption in its favor. Thus, even though contention
over the very idea of evolution remained in the general culture, the project
of natural theology receded from view.

In recent times, however, there have been attempts to revive natural the-
ology based on new notions derived from physics (the anthropic principle)
and from biochemistry (irreducible complexity). It is the latter approach,
advanced by Michael Behe and a number of other “intelligent design the-
orists,” that most directly attempts to recover Paley’s argument and that
in consequence will concern us (Behe 1996, 2001).
2
Behe summarizes the
functional complexity at the molecular level of a variety of biological phe-
nomena, such as the clotting of blood, the locomotion of bacteria, the im-
mune system, the biochemistry that underlies vision, and even the origin
of life. He concedes early on that he accepts natural selection and com-
mon descent; he even concedes that there are plausible accounts of how
the vertebrate eye could have evolved (see, for example, Futuyma 1998;
Gilbert 2000). He thus bears witness to the changed presumptions in favor
of selectionist explanations. Having done so, however, Behe shifts the ar-
gument to the greater complexity and functional intricacy of the molecules
that underwrite macroscopic biological adaptations. How possibly, he asks,
could this particular biochemical feature or trait, which requires X num-
ber of components, each with a precise structure and function, have arisen
by natural selection, when a loss or defect in any one of the components
must cause a loss of the function of the trait? Any system, that is, with X−1
components will have no function, no fitness – nothing upon which nat-
ural selection can act. Indeed, Behe claims that “many biochemical sys-
tems cannot be built up by natural selection working on mutation; no di-
rect, gradual route exists to these irreducibly complex systems” (Behe 1996,
p. 202).
P1: KAF/KAA P2: KaF
0521829496c09.xml CY335B/Dembski 0 521 82949 6 March 10, 2004 1:5
Darwinism, Design, and Complex Systems Dynamics

175
Behe’s argument from incredulity takes the following schematic form:
(1) Natural selection has to be gradual, linear, and sequential if it is to
result in adapted traits.
(2) Observed molecular functional complexity of Y is inconsistent
with (1).
(3) Natural selection cannot possibly account for Y.
Behe tacitly assumes that the molecular components of cells, if not cells
themselves, are little machines, artifacts with moving parts. Since the an-
swer to the question of whether natural selection could possibly have been
the causal explanation of the appearance of a particular adaptive trait is
negative, and since Behe denies that there is any other possible natural ex-
planation, he reasons by disjunctive syllogism that the only other possible
explanation is intelligent design. We can call this “the argument to the only
alternative.” The rhetorical advantage of this strategy is that, having elimi-
nated the only other possibility, it is not necessary to provide any positive
argument for design. But the strategy will persuade only if three conditions
are fulfilled: (1) there are no empirically adequate answers to the ‘how pos-
sibly’ questions Behe poses; (2) cells and their molecular components are
no more than machines; and (3) there is a simple dichotomy at the level
of possible explanations. We will consider Behe’s argument in the light of
these three conditions. We will find reason to doubt whether any of them
is strong enough to sustain Behe’s argument.
2.
the argument against incredulity
When considering ‘how possibly’ questions, one needs to think about the
kinds of systems in which changes occur over time (i.e., the nature and dy-
namics of these systems). Intelligent Design (ID) theorists, such as Behe,
presuppose that the only possible systems are ones that are highly decompos-
able. This is implicit in the strong analogy they draw between biological or

biochemical systems and man-made machines; the latter presume a linear as-
sembly model of systems. Unfortunately, contemporary hyper-adaptationist
versions of Darwinism – arguments that look for an adaptationist explana-
tion for virtually every trait or evolutionary phenomenon – meet ID on the
same ground by employing similar classes of models (Dawkins 1986; Dennett
1995).
3
In effect, the ID theorists and the hyper-adaptationists are taking in
each other’s laundry when they both use the terminology of design, even
as they differ in the source of the design. This way of framing the debate
between evolutionists and creationists foreshortens the space of what will
count as possible explanations. Unfortunately, too, this is something both
camps seem to prefer. Both assume that there are only two alternatives and
P1: KAF/KAA P2: KaF
0521829496c09.xml CY335B/Dembski 0 521 82949 6 March 10, 2004 1:5
176
Bruce H. Weber and David J. Depew
argue that the other alternative is impossible. A knockout blow is the aim of
both a Dennett and a Behe, a Dawkins and a Dembski.
But deploying ‘how possibly’ questions in order to foreclose any but one
answer to ‘why actually’ questions is a very risky business. Because ‘how pos-
sibly’ claims seldom, if ever, reach the status of ‘why necessarily’ answers
in favor of the disjunct they prefer, they can be upended by plausible ‘why
actually’ hypotheses that put the burden of proof back onto the questioner.
Suppose for a moment that ‘why actually’ scenarios or hypotheses are given
as plausible answers to questions posed by an ID theorist about the origin of
a particular biological mechanism. Then either (1) a retreat from the claim
of the ‘irreducible complexity’ of that trait must be made, or (2) it must
be acknowledged – especially if the number of such retreats accumulates –
that the assumption of decomposable systems acted upon by gradual, linear,

sequential selection is not valid. Both eventualities weaken the ID case. If
evolutionary theory is merely a stalking horse for materialism, then ID argu-
ments do have a certain rhetorical point. However, the problem is that ID
theorists foreclose the ‘how possibly’ question too quickly, slyly converting
the ‘why actually’ question into a dichotomy of either external design or
selection.
What if natural systems are not strongly decomposable but in fact are gen-
erated by and composed of parallel processes? What if such natural systems
can “limp along” when they open new survival space with the beginning
of the emergence of novel properties, and what if these properties can be
polished by selection to a shine of apparent design? Such scenarios, if bio-
chemically plausible, could take the force out of the appeal to irreducible
complexity and the argument from incredulity. That is why well-documented
processes, such as gene duplication, are relevant. They allow for parallel and
divergent evolution of enzyme functions that can open new energy sources,
produce new metabolic pathways and complex functions, and cope with
potentially lethal mutations (Shanks and Joplin 1999; Thornhill and Ussery
2000; Deacon 2003).
Behe argues that “the impotence of Darwinian theory in accounting for
the molecular basis of life is evident ...from the complete absence in the
professional scientific literature of any detailed models by which complex
biochemical system could have been produced” (Behe 1996, 187). Yet liter-
ature published before Behe wrote, and subsequent papers in the profes-
sional scientific literature, have provided just such ‘how possibly’ scenarios;
and further research in some cases is shifting to the ‘why actually’ issues.
For example, a chemically plausible route for the evolutionary origin of
the Krebs citric acid cycle has been proposed. It might serve as a more gen-
eral model for the emergence of complex metabolic pathways (Melendex-
Hevia, Waddell, and Cascante 1996). Behe could respond to such arguments
with the admission that this particular system, the Krebs cycle, has turned out

not to be irreducibly complex, since an explanation of its emergence can be
P1: KAF/KAA P2: KaF
0521829496c09.xml CY335B/Dembski 0 521 82949 6 March 10, 2004 1:5
Darwinism, Design, and Complex Systems Dynamics
177
plausibly advanced. Behe has taken just such a position about the emergent
phenomena of patterns of chemical reactions in the Belousov-Zhabotinsky
(BZ) reaction, which was originally developed to model the functional com-
plexity of the Krebs cycle (Belousov 1958; Behe 2000; Shanks 2001). It turns
out that the BZ reaction can be understood in terms of theories of nonlin-
ear chemical kinetics and nonequilibrium thermodynamics, even though
the emergent phenomena were not anticipated by the decomposability as-
sumption of conventional chemical descriptions (Tyson 1976, 1994). The
BZ reaction can be used as a model to explore the dynamics of other com-
plex systems, including biological ones (Shanks and Joplin 1999; Shanks
2001). Behe does concede that the origin of some biochemical systems may
also be explainable by natural processes, and hence by processes not fitting
his definition of irreducible complexity (Behe 2000). But wait a minute.
We were assured by Behe that no complex biochemical system, including
metabolic pathways such as the Krebs cycle, had been given a Darwinian
explanation in response to his ‘how possibly’ questions (Behe 1996). There
are many classes of Darwinian explanations, and Darwinian explanations
are only a subclass of naturalistic explanations. All we need at present is a
plausible naturalistic explanation, whether Darwinian or not. As similar ar-
guments are extended to other metabolic pathways, however, the scope of
the phenomena that Behe claims to explain by ID shrinks. Well, Behe will
say, even if metabolism is capable of naturalistic explanation, still there are
other very complex phenomena that surely resist explanation. But at a cer-
tain point, it is this very assertion that is in dispute.
Consider the following case, in which explicitly Darwinian naturalism

figures. Behe has claimed that “no one on earth has the vaguest idea how the co-
agulation cascade came to be” (Behe 1996, 97, emphasis in original). But there
are, it turns out, plausible scenarios by which crude, primitive blood clotting
could have occurred in organisms that employed proteins from other func-
tions to achieve a marginal stoppage of bleeding (Doolittle and Feng 1987;
Xu and Doolittle 1990; Doolittle 1993; Miller 1999). Through gene dupli-
cation and shuffling of the subregions of genes coding for domains within
the proteins produced, such protein functions could expand, diverge, and
become targets of selection that could lead to improved clotting. Proposals
such as this can be tested by making predictions about both the specific
relationships of sequences of the various genes within the pathway and to
sequences of genes for proteins in more primitive organisms that might
have been the origin of the original genetic information. This kind of work
is proceeding and is beginning to provide the basis for shifting to the ques-
tion of ‘why actually’ such a complex cascade has actually arisen in evolution
by natural selection.
Again, Behe asserts that “the only way a cell could make a flagellum is
if the structure were already coded for in its DNA” (Behe 1996, 192, em-
phasis added). The bacterial flagella that Behe considers have over forty
P1: KAF/KAA P2: KaF
0521829496c09.xml CY335B/Dembski 0 521 82949 6 March 10, 2004 1:5
178
Bruce H. Weber and David J. Depew
protein components, and it would seem that he has a strong case here. Yet
only thirty-three proteins are needed for fully functional flagella in some
bacteria not considered by Behe. Even so, such a number constitutes an
explanatory challenge. Clues to how such a system might have evolved, em-
ploying processes of protein “recruitment” and gene duplication followed
by divergence, are provided by the roles of ion gradients in the organiza-
tion of bacterial metabolism and their possible role in self-organizational

and emergent phenomena, a subject to which we will return (Harold 1991;
Harold 2001).
Another example Behe uses is the truly complex vertebrate immune sys-
tem of over ten thousand genes. He claims that “the scientific literature has
no answers to the origin of the immune system” (Behe 1996, 138). But shortly
after Behe published his book, the discovery of the RAG transposases and
transposons in contemporary vertebrate immune systems suggested possi-
ble routes by which “reverse transcription” (similar to the way in which
the HIV virus converts its RNA message into DNA and incorporates it into
T-cells) could have worked, in addition to gene duplication and domain
shuffling through RNA splicing mechanisms, to generate the immune sys-
tem (Agrawal, Eastman, and Schatz 1998; Agrawal 2000; Kazazian 2000).
There is, to be sure, a long way to go in developing our understanding of
how immunity came into existence. But it should be clear that the reason-
able way to proceed is through pursuing the combined research programs
of molecular genetics, genomics, proteomics, and informatics rather than
simply by declaring the immune system to be too complex for any other
than ID explanations (if indeed they can be called explanations). In time,
Behe may be forced to concede that the immune system, as well as blood
clotting and metabolic pathways, are subject to naturalistic explanations
and, although complex, are not in the end irreducibly complex. We are just
now entering a new age in understanding the relationship between genes
and biochemical pathways and systems. Why does Behe think that, just as
this new cascade of research gets under way, he can put a stop to the whole
business?
Finally, let us return to Paley’s example of the eye. Behe concedes, as men-
tioned earlier, that the eye as an organ might have a Darwinian explanation.
But he argues that the biochemistry of vision is too complex to have any
other than a design explanation. What can be said about this? Photosensi-
tive enzyme systems exist in bacteria, functioning in energy metabolism, and

indeed may be quite ancient (Harold 1986; Nicholls and Ferguson 1992).
Proteins that function in the lens turn out to be similar to the sequences
of amino acids in heat shock proteins, and to various metabolic enzymes
that function in the liver (Wistow 1993), suggesting once again a process
of parallel processing via gene duplication and divergence as a new function
becomes selectively advantageous. More dramatically, recent developments
in our understanding of the molecular basis of biological developmental
P1: KAF/KAA P2: KaF
0521829496c09.xml CY335B/Dembski 0 521 82949 6 March 10, 2004 1:5
Darwinism, Design, and Complex Systems Dynamics
179
processes reveals that certain genes responsible for the development of eyes,
such as the pax 6 gene, arose only once in the history of life and, although
not coding for eyes per se, code for the processes that lead to the construc-
tion of eyes in various lineages (Gehring 1998; Gilbert 2000, 705; Carroll,
Grenier, and Weatherbee 2001). Comparisons of the genes regulating de-
velopment in different species hold the promise of giving us clues to the
origin and evolution of complex, multicellular organisms and their organs
(Carroll, Grenier, and Weatherbee 2001).
A recurring theme in the specific research areas just mentioned is that
evolutionary processes occur through parallel processing rather than by se-
quentially adding one perfected component to another. Thus evolutionary
biologists, other than those committed to an extreme form of adaptationism, have
explanatory resources that ID theorists deny (see, for example, Thornhill
and Ussery 2000). (That is why ID creationists and hyper-adaptationists have
a common interest in narrowing what counts as evolutionary theory and as
Darwinism.) Even if organisms are conceived as decomposable artifacts – a
position against which we will argue later – they are not analogous to the
assembly protocols for watches, in which the components are specifically
designed and perfected for unique functions and then put together in a

specified sequence. Interestingly, engineers now find that in designing complex
machines (such as a processor composed of over forty million transistors),
it is more efficacious to have the components combined in a functional
pattern than to assemble them from perfect components. Indeed, the best
strategy is to get as quickly as possible to a crudely functional whole and then
to remove and replace less reliable subsystems. Challet and Johnson estimate
that one-half of the components could be significantly imperfect and yet the
system as a whole would function reliably (Challet and Johnson 2002). With
complexity comes redundancy and parallelism that can give functionality;
and with functionality comes pressure for improved components over time.
Thus, even allowing ID theory its artifact metaphor, it poses its ‘how possibly’
arguments the wrong way around. The function of the whole system does not
depend upon perfectly designed and articulated components.
4
The questions that need to be addressed, then, are what types of systems
are living organisms – and ecosystems as well? What are the dynamical prin-
ciples of such systems? Do design and selection exhaust the possibilities of
explanatory space in such systems, or are there other relevant natural pro-
cesses, such as self-organization? To these issues we now turn our attention.
3.
self-organization and emergence in complex
natural systems
Both ID and hyper-adaptationist Darwinian (HD) theorists share assump-
tions about the nature of organisms and systems. These shared assumptions
imply that self-organization is not a serious contending explanatory resource